Ovarian cancer has the highest mortality rates of female cancers. The underlying biology of this disease is poorly understood and the existing approaches for diagnosis and prognosis are inadequate. Based on the hypothesis that methylation changes at CpG islands play a prominent role in cancer, we developed a microarray-based technique called differential methylation hybridization (DMH). DMH allows for the global analysis of CpG island methylation in tumor genomes, a method to determine methylation profiles of cancer cells, and has the potential use as a site specific diagnostic approach to test whether demethylation at specific loci can be identified. Previously, we used DMH on a small group of patients to perform methylation profiling of ovarian cancer, establish proof-of- concept and lay the foundation for genome wide screening of methylation to examine epigenotype-phenotype relationships in ovarian cancer. In the present study, we will perform methylation profiling of ovarian cancer using DMH arrays containing 21,000 CpG islands. Methylation profiles of ovarian tumors from patients diagnosed with early and advanced ovarian cancer will be compared to normal ovarian surface epithelium. Overall methylation patterns will be used as an "epigenetic signature" to characterize specific types and stages of ovarian cancer, and these molecular signatures will be correlated with clinicopathological parameters of the patients. Specific methylation patterns identified can later be applied to predict patients' outcome in clinical settings. The resulting array data will be used to identify the specific CpG island sequences frequently hypermethylated in ovarian cancer. As these are not normally methylated in adult tissues, CpG island methylation represents one of the most prevalent tumor specific markers yet identified. When associated with specific genes, CpG island methylation may have consequences for ovarian tumor types. In addition, we have further refined DMH by using expressed CpG island sequence tags (ECISTs) for dual detection of CpG hypermethylation and gene expression/silencing in cancer cells. ECISTs exist in the genome, and their GC-rich fragments can be used to screen aberrantly methylated CpG sites in cancer cells. The exon-containing portions can be employed to measure levels of gene expression simultaneously. Using an ECIST panel we have recently developed, we will identify hypermethylated loci and at the same time confirm their association with gene silencing in the ovarian cancer samples. This approach will also allow us to study gene promoter activity in ovarian cancer and assess the importance of screening for gene promoter functions in this disease. In summary, this study will address the clear need for developing better tools for the screening and staging of ovarian cancer, as well as the need to identify new markers that adequately address the complexity of this disease. Methylation profiling of ovarian tumors could provide a more focused test for reactivation of methylation-silenced genes as therapeutic targets and thus play a role in the rational basis for new clinical strategies designed to alter this fundamental process in ovarian cancer.